I can confirm similar issues with IPv6 stack enabled on a Balance Two router with a bunch of Peplink 24 Rugged Switches supplying several AP AC One Minis and AP One AX Wifi Access Point connected to it via PoE.
We are facing particularly problems invluding the one with DNS packets getting lost via WIFI APs from different VLANs as the issues described in the post of the link you provided.
We are not using Starlink, but Vodafone Business Cable WAN connection and are connected to a Fritz!Box 6660 Cable (BTWO → DMZ).
The FB delegates suitable IPv6 sub addresses from our fixed IP range to the BTWO and its clients (Prefix Delegation).
We also had DNS over Https enabled using IPv4 DNS Servers from nextdns.io all the time testing. (Question: I am not sure if this leads to conflicts if IPv6 stack is active?? Unsure if I then should insert the IPv6 DNS servers instead? Does anyone know about the right setting and can please give advice?). Then I disabled that at the end, but without any noticeable improvement.
However clients (mobiles, IOT devices, Homepods) are facing delays while web browsing or webpages or streaming servers are not reachable or only after several tries.
Additionally, our clients are facing issues to reach devices connected to other VLANs, for which access protection was configured only using the IPv4 LAN address via BTWO‘s firewall rules (Note: IPv6 firewall rules… how should that work with ongoing changing local IPv6 fd fe addresses of the clients? 
)…
I had already discovered further issues with Peplink IPv6 and VLANs, see here:
https://forum.peplink.com/t/ipv6-support/3675/42?u=ckirch
I ended up disabling IPv6 on the BTWO to reduce these complaints.